Is AI the ‘Silver Bullet’ for Starmer’s Energy Ambitions?

“The national grid has become the single biggest obstacle to the deployment of cheap, clean power generation and the electrification of industry”, claimed Sir Keir Starmer’s government in their manifesto. By now we are all familiar with Labour’s ‘Plan For Change’, outlining the five key missions of this Parliament. The final one of these (‘Make Britain a clean energy superpower’) promises to deliver clean power by 2030, citing economic, national security, and environmental incentives for this ambitious target. The subsequent AI Opportunities Action Plan proposes AI as the single biggest tool that could be leveraged to achieve this goal – and justifiably so, with AI offering a host of solutions to age-old grid issues. Unfortunately, as ever, it’s not as simple as asking ChatGPT to design a new energy system for us. If we want to genuinely advance net zero goals whilst furthering Starmer’s dream of becoming a leader in the global AI arms race, we need to seriously address how these technologies are being deployed.

‍

How Can AI Help?

AI holds enormous potential to advance the push for a more sustainable energy system in the UK. Predictive models can be used to more accurately forecast renewable energy sources, optimise the efficiency of both electricity generation and storage, improve energy load forecasting, and flag maintenance needs earlier than previously possible. Exciting news for those who have spent decades pondering how best to tackle the ‘Energy Trilemma’ (how to ensure security of supply, decarbonise generation, and lower costs for consumers), as it seems the historical incompatibility of this trifecta could soon be an issue of the past. Of course, some teething problems persist, such as barriers to accessing data for training new models, the need to upgrade older infrastructure, privacy and security concerns, and a skills gap in the workforce. However, these are slowly but surely being tackled through a range of innovation grants, education and upskilling programs, evolving data protection laws, and a steady stream of investment. What is more concerning, ironically, is the AI-induced increase in pressure on the very same, already-creaking national grid.

‍

‍

One of the key roadblocks in the transition to clean energy is the enormous volume of applications waiting to be connected to the UK grid. With projects facing wait times of 10 to 15 years, and 63% of applications not having changed status between 2018 and 2023, the historical ‘first-come, first-served’ approach to connection has led to a queue of 701 Gigawatts. This is enough energy to power every single home in the UK four times over. Reducing this wait time to 6 months would not only allow the harnessing of renewable energy projects currently stuck in the queue, but could also save up to £75 billion over the next decade. AI has been proposed as a natural solution, with powerful matching algorithms that could make the connection process more efficient.

‍

What’s Stopping It?

However, such AI models are also run on electricity. Specifically, they rely on power-hungry datacentres, many of which have committed to net-zero or even fully renewable energy sources in the face of rising bad press about their fuel consumption. According to John Pettigrew, the CEO of National Grid, the electricity demand of such data centres is expected to increase sixfold over the next decade – a jump almost entirely attributable to AI usage. These data centres must also join the queue to connect to the grid, leading to an unfortunate catch-22 whereby AI falls victim to the very problem it has been employed to solve.  

‍

‍

Labour seems to be cottoning on to this paradox though. The government has recently announced the creation of an AI Energy Council, led by Energy Secretary Ed Miliband and Technology Secretary Peter Kyle. The hope is that this council will work with energy companies to gain further understanding of the unique nature of AI’s energy demands and challenges, to help strategise about its development more effectively. Data centres have also been upgraded to being a Critical National Infrastructure (alongside energy and water systems), meaning they will receive greater government support for critical incidents, as well as the formation of a team of senior government officials to support the sustainable development of the industry. Progress may be slow, but these are some promising steps in the right direction.

‍

AI is developing at scale, whether the energy grid is ready for it or not. If the government is smart (and careful) about the manner of its development and deployment, it could simultaneously accelerate the transition to a greener national grid, whilst positioning itself as a global AI leader, neatly checking the boxes of two of its major policy goals. The future is bright if Starmer’s government can get AI right – but execution of this complex and ambitious project remains to be seen.